1. Detecting Computer Intrusions Using Behavioral Biometrics Ahmed Awad E. A, and Issa Traore University of Victoria PST’05 Oct 13,2005

2. Motivation Attacks targeted by intrusion detection systems can be divided in three forms: user-level, system-level, and network-level. Typical user- level attacks consist of masquerade attacks. Even though a typical computer intrusion involves exploiting together several of these vulnerabilities user-level attacks remain one of the most recurring forms of intrusions because successful user-level attacks always serve as prerequisite for most forms of intrusions.

3. Motivation We propose in this work a new approach to user profiling based on biometrics. The profiles computed in this case are more accurate than those obtained through the traditional statistical profiling techniques, since they are based on distinctive biological characteristics of users. The utilization of biometrics technology, however, has so far been limited to identity verification in authentication and access control systems. Hence important security applications such as intrusion detection systems have been left out of this technology.

4. Approach In this work we combine a new behavioral biometrics based on computer mouse dynamics and enhanced version of keystroke dynamics biometrics. Mouse and keystroke dynamics biometrics fulfill all the characteristics required for intrusion detection since it allows passive, dynamic, and real-time monitoring of users Our aim is to be able to construct a biometric signature which can be used to identify the user and detect possible intrusions.

5. Detector Architecture Client module, runs on the monitored machine (e.g. potential victim), is responsible for mouse movement and keystroke data collection. Security administrator makes sure for local users that the data collection software is installed on local machines. Server module is in charge of analyzing the data and computing a biometric profile. Computed profile is then submitted to a behavior comparison unit, which checks it against the stored profiles.

6. Detector Architecture

7. Mouse Dynamics monitor all mouse actions generated as a result of user interaction with a graphical user interface process the data obtained from these actions in order to analyze the behavior of the user. The behavioral analysis unit utilizes neural networks and statistical approaches to generate a number of factors from the captured set of actions these factors are used to construct what is called a Mouse Dynamics Signature or MDS, a unique set of values characterizing the user’s behavior over the monitoring period.

8. Active profiles of a given user compared to the reference profile of a different user

9. Active profiles compared to the reference profile for the same user

10. Average Speed for Different Types of Actions, comparing large number of sessions for two different users

11. Types of Actions Histogram, comparing large number of sessions for two different users

12. Keystroke dynamics measure the dwell time (the length of time a key is held down) and flight time (the time to move from one key to another) for keyboard actions. collected actions are translated into a number of digraphs or trigraphs. our detection algorithm generates a Keystroke Dynamics Signature or KDS, which is used as a reference user profile and matched against active user profiles to dynamically detect masqueraders.

13. Di-graph approximation matrices for two different users

14. Experimental Results 22 participants 9 weeks divided the participants into 2 groups: a group of 10 representing authorized users and a group of 12 representing unauthorized users. false negative rate of 0.651% false positive rate of 1.312%

15. Simulated Attack: one insider and two outsiders masquerading as a legitimate user MDH: Direction of Movement histogram ATH: Type of Action Histogram ATA: Average Movement Speed for Action Types MSD: Movement Speed compared to Traveled Distance

16. BioTracker BioTracker provides an automated method of authenticating or verifying an individual based upon physical or behavioral characteristics. A prototype has been implemented and tested in the ISOT lab. PCT patent application with the World Intellectual Property Organization filed by the University of Victoria (PCT/CA2004/000669). Web Site: http://www.isot.ece.uvic.ca/projects/biotracker/http://www.isot.ece.uvic.ca/projects/biotracker/ E-Mail: isot@ece.uvic.caisot@ece.uvic.ca